Superpedestrian’s smart scooters use AI to identify potential problems

Make way for a hefty e-scooter that can take care of itself. 

After five years of development, Superpedestrian, the Cambridge, Massachusetts-based transportation robotics company, is finally ready to ship its “smart” electric scooter to rental companies. 

If it sounds at all familiar, that’s because it is. Last year, Mashable had a chance to demo the company’s prototype e-scooter and companion Vehicle Intelligence platform. Now, that autonomous maintenance system and the beefy two-wheelers it lives inside are ready for the streets. You won’t be able to buy the device for yourself. But, since Superpedestrian already has deals set with a few clients, chances are you’ll stumble across them when you rent an e-scooter starting next year.  

New e-scooters, however, are not Superpedestrian’s endgame. Assaf Biderman, the company’s CEO, would much rather focus on building the electronics powering e-scooters and the control systems that allow fleet operators to efficiently monitor and fix vehicle problems. But, as he explained at a San Francisco demo last week, what’s currently out there doesn’t cut it. So, the company took matters into its own hands and developed a proof-of-concept e-scooter to show how its software could reduce, or even eliminate most maintenance issues. 

A broken e-scooter, for example, can be offline and unavailable on a rental app for up to 19 days before repairs can be completed. That’s a figure Superpedestrian wants to bring down to just a few hours. 

For a good snapshot into the current state of e-scooter management, consider this recent Sifted interview with Bolt CEO Markus Villig. Villig, whose Estonian ride-sharing and food delivery company will be making its way to the U.S. soon, revealed that his e-scooters last for only about five months and aren’t profitable — a dilemma that’s all too familiar to most other scooter operators. 

It’s precisely those issues that Superpedestrian hopes to solve. To that end, the company trained its e-scooter AI to detect failures and predict when something’s likely to go wrong. Sensors laid throughout the e-scooter allow it to monitor and instantly report any potential issues with the batteries, motor, brakes, or wheels. If, say, an e-scooter detects water or cooling problems in its battery, it’ll disconnect itself, and flag the system for a check-up. A technician can then go to the scooter and decide if a more intensive repair is needed, or if a simple battery swap will do.

“We give the vehicle autonomy to decide if it’s safe,” Biderman says about the e-scooter’s distinguishing feature.

Thanks to its self-diagnostic software, Superpedestrian expects its e-scooters to last for more than 2,500 rides each, which is currently way more than any of the current crop littering our streets. And with a bigger, longer-lasting battery rated for up to 60 miles of riding, these e-scooters only need a charge once every three to seven days, instead of daily. 

Eventually, Superpedestrian plans to specialize in software only, giving e-scooter operators a way to track battery health, brakes, e-scooter distribution, rider speed, and more. For now, the company is content with its recently announced $20 million in new investments, a financial cushion that’s sure to ease the rollout of its smart e-scooter fleet.